|
No renovation beyond the most superficial should be attempted without a sound under standing of structural principles. One enthusiastic renovator we know did a beautiful remodeling of his parents’ basement recreation room which included the removal of an ugly Lally column that stood in the center of the space. It did not take long for him to recognize his structural error; luckily his parents were not home at the time of the collapse. The discussion below will not qualify you (by any stretch of the imagination) to practice engineering (which would be foolish and illegal). It will, hopefully, provide you with some basic structural tenets and some idea as to when to consult a professional. We have found that most people have only a vague idea of what makes a building stand up. Most people know (with the exception of the young man cited above) that you can’t remove a “structural wall” without dire consequences, but few laypeople have any idea of how to identify a structural wall from one that's non-load-bearing. Many people we have talked to are under the erroneous impression that only the outer walls of a building are structural and that all of the inner partitions* merely divide the space into rooms. Others, on the very conservative side, think that all partitions are structural and refuse to consider the removal of anything that resembles a wall. The structure of a small house or apartment building employs the same basic principles as the megastructure of a skyscraper. In the attempt to gain insight into how any building works, it's more valuable to think of the various structural components as working together to transfer loads downward to the ground rather than to consider them as holding something up. When dealing with buildings, one of the essentials to keep in mind is structural continuity. This kind of continuity applies to human anatomy as well. Our body skeleton carries the weight of our various organs and tissues from one bone to the other until the entire load is transferred through our feet to the ground. Similarly, the loads on a building, both inherent (dead loads: the weight of materials) and applied (live loads: people, furniture, and snow), must be transferred through the various horizontal, vertical, and diagonal structural members to the foundations and from there to the ground. A building consists of three main structural systems that interrelate to transfer the loads downward. The horizontal system includes the floors and flat roof which directly receive the weight of people, furniture, snow, and rainwater and transfer that weight to the vertical supports. The vertical supports are either columns, walls, or partitions that transfer the loads vertically to the foundations. The foundations serve two purposes: to transfer the building loads to the ground and to anchor the building. (A diagonal roof transfers loads both vertically and laterally; this presents its own set of problems.) We will use the term “wall” for the divider between the interior of the building and the outside. A “partition” is de fined as a floor-to-ceiling divider separating one interior space from another. Both walls and partitions may be either structural or nonstructural. VERTICAL SUPPORTS There are a number of ways to structure a simple building: you can frame it with columns and beams, you can use load-bearing walls to support the floors, or you can construct the building using light-frame construction methods, which in essence combine the other two systems. The Empire State Building in New York is of steel-frame construction. In that structure and in thousands like it of smaller stature, the loads of the building are transferred through the steel frame downward, much as a person’s weight is carried to the ground by the body’s skeleton. The floor loads are transferred through beams to girders to columns and down through the columns to the foundations. The walls act to enclose the building, are non- structural, and are referred to as curtain walls. A small house could be constructed using exactly the same principles if framed with lightweight steel components or timber, a method known as post-and-beam construction. Most high-rise apartment buildings (more than six stories) and many shorter multi-residences are constructed by means of this point-support system using either steel or reinforced-concrete columns and beams. In geographic areas where earthquakes are a threat, walls, floors, and foundations must be designed to resist these lateral vibrations. Seismic design isn't discussed in this guide. A small warehouse with masonry walls is structured on the principle of the load-bearing wall. Some of the walls and partitions of this building do more than enclose space for reasons of privacy and the need to hold heat in winter. These walls and partitions transfer the weight of the floors and roof to the foundation all along the length of the walls. Since the bearing wall or partition acts as a unit to transfer the loads, there is a certain limit to the number of openings (such as windows and doors) that can be placed in it (Ill. 4). When the penetrations in a wall become so large that the wall can no longer act structurally as load-bearing, other arrangements are needed for the transfer of the weight. In such cases, the loads are distributed through lintels (acting as beams), which transfer them to point supports— that's , columns. Many small apartment buildings and urban row houses are based on the load- bearing wall system and so is the solid-brick or stone single-family house. Some very old apartment houses are hybrid in construction in that most of the structure is carried by load-bearing walls with the exception of a few rows of columns placed in locations where bulky walls are undesirable. In a building of light-frame construction, the house is framed using slim structural members spaced closely together. The walls and partitions consist of studs (mini-columns) spaced 12”, 16”, or 24” apart instead of heavier columns that might be spaced 6’ or 8’ apart in wood post-and- beam construction, 12’ to 18’ apart in a concrete- frame building, or even farther apart in a steel-frame building. By positioning the studs so close together and connecting them with sheathing for bracing (usually sheets of plywood), the wall is converted into a load-bearing wall. If all the walls and partitions are constructed in this way and the penetrations for the doors and windows are limited and small (narrower than 6’), there is no need for any columns to support the floors or roof above: If the interior space is large, or if the partitions are inconveniently placed for load-bearing purposes, columns can be used to support the structure above, making the building a hybrid of light-frame and post-and-beam construction methods. Most single- family houses with wood siding and many houses faced in brick utilize light-frame construction. So do many small suburban row-house complexes. In the vertical support system of the house, especially in walls and short columns, loads are transferred mostly through compression—that is, the molecules or fibers in the column are being com pressed or squeezed together. Most structural materials are strong in compression, meaning that they can carry a great deal of load while being squeezed, but to varying degrees are weak in tension (stretching). Both of these stresses occur when a structural member is positioned horizontally. When a beam is placed between two supports and is loaded, it's said to be subjected to “bending” stresses (a combination of compression and tension). The beam gives, or deflects, under the weight. If we concentrate on what is happening somewhere in the middle of the deflecting beam, we will see that the uppermost molecules in the member are being compressed, whereas the lower molecules are being stretched, or placed in tension. The tension stresses are greatest at the bottom of the section and de crease toward the center. The compression stresses are greatest at the top of the beam and decrease toward the center. The centermost molecules (those running along an imaginary line called the neutral axis) aren't subjected to any stresses at all. Because of the dual stresses (tension and compression) that occur in horizontal structural members when subjected to bending, they require different structural analysis than do short vertical supports, which are generally subjected only to compressive stresses. HORIZONTAL SUPPORTS Let us suppose that a hiker has to cross a stream that's 4’ or 5’ wide. He would know that a long twig about 2” in diameter would not support his weight across the gap. (We are assuming, of course, that he is a tightrope walker.) He would probably decide that a tree trunk 4” in diameter would do the trick. Were the gap increased from 5’ to 15’, he might question whether a 4”- diameter log (assuming he could find one long enough) would be able to support his weight across the longer span. Presented with this problem, he would probably turn around and look for a log thicker than 4”. He suspects, correctly, that the 4” log would bend or perhaps would even snap in two if he attempted to walk on it. Whether we realize it or not, most of us have some basic knowledge of the tenets of structural design. In the example just given, the hiker knew that the dimensions of the structural member required to bridge the gap were dependent on the length of the span and the lo4ds to be supported. In fact, even highly sophisticated structural problems that involve spanning between supports can be reduced to four variables: 1. The span—the distance between the supports and how the member is connected to its supports. 2. The loads to be carried—how much weight, where it's concentrated, and for how long. 3. The cross-sectional dimensions of the structural member—the thickness and shape of the cross section of the member. 4. The material being used—its strength, characteristics, ability to withstand the stresses it's being subjected to. The Span The word “span” means the distance between supports. The most common span in residential architecture is the “simple span.” This refers to a horizontal member (a plank, joist, beam, or girder) that's supported at both its ends by either walls or columns. The ends may be secured to the supports in some way, but the connection isn't rigid enough to prevent slight movement (or rotation). When a simple span is loaded (that is, when weight is put on the beam), it tends to give somewhat under the weight. This giving, or bending (the combination of tension and compression), causes the beam to become slightly distorted. We use the term “deflection” to describe the amount of distortion. In the case of the hiker, the distance between the supports was in creased from 5’ to 15’ without any increase in the thickness of the cross section of the log or in the load that it would have to bear. This increase in the span would cause the member (the log) to deflect even more (under its own weight). Were you to further increase the span (without changing the cross-sectional dimensions of the spanning member), the beam would eventually fail. Since these variables (the span, the loads, and the cross-sectional dimensions of the beam) are interdependent, an increase in one of them (in this case, the span) would require either an in crease in the cross section of the member or a reduction of the load. Supposing that these alter natives are impractical, the obvious thing to do is divide the long span into shorter spans. This can be done in a number of ways. The first solution is to erect additional supports and to bridge the spans between the supports with short beams. These simple beams, with non-rigid joints, are most often seen in wood-frame construction. The second method is to use one continuous beam across all of the supports. This would reduce the deflection in any one span and would be the more efficient solution. Another way to obtain the structural benefits of a continuous beam without the problems of importing an extraordinarily long girder to the construction site is to be sure that the joints that tie the beams to the columns and to each other are rigid (non-rotating). The simply spanned beam, one that spans from one support to the other, is commonly used in residential architecture and is perfectly adequate in most cases. We don’t often see long continuous beams in ordinary house construction. Very long milled sections are unusual and expensive; the longest section you are likely to find without special ordering would be about 22’. Laminated beams (manufactured structural members that are produced by pressure-gluing short, thin wood strips together) are expensive both to purchase and to move to the site. Likewise, rigid connections aren't often used in wood-frame construction. It is possible, however, to build up a long continuous beam out of lengths of 2” -thick lumber. (See Section 29, Ills. 3 and 4.) All welded steel connections are considered rigid connections, and any poured-in-place concrete frame is by definition a rigid frame. THE CANTILEVERED SPAN: There are in stances where the design of a building calls for an overhang or balcony, eliminating the end support of the beam. In this situation the beam is cantilevered. The structural member will have two sup ports, but only one of these will occur at the end. The other support occurs somewhere along the length of the beam. When the end of the cantilevered portion is loaded, the effect is like a seesaw. The downward motion at the cantilevered portion must be opposed by a downward force at the beam’s opposite end. Otherwise, that end will go up like the end of a seesaw. A wall or other similarly heavy weight applied at the non-cantilevered end will act to balance the beam. Otherwise, the connection between the beam and its end support will have to be made very strong. LOADS A beam’s primary function is to carry the loads applied along the length of its span to its sup ports. Three crucial questions about the loading of a beam must be answered: How much weight must the beam carry? For how long? Is the load concentrated at any one point or is it evenly distributed along the beam’s length? If the load is evenly distributed along the length of the span, every inch of the member is put to work supporting the weight applied to it, in addition to transferring the weight to the nearest support. The same total load applied at a single point causes different problems. If the load is applied at or near midspan, the bending and consequent deflection will be greater. The length of time the load will be sustained is an important consideration as well. A wood member might be fine to support an enormously heavy load for a few minutes or even as long as a week. It will bend a bit under the load and deflect to some extent, but if the load is removed it will come back to its original shape. If the load is kept there for as long as a year, strange as it may seem, the material gets fatigued. When the load is removed, the member will not assume its original shape. The magnitude of the loads to be carried is the decisive factor in the design of a beam. The greater the load on a beam, the greater the bending stresses, and the greater the danger of the beam failing. If the loads are too great for the beam, it will fail. Presented with a large load and a fixed span, the only variable left is the cross-sectional dimensions of the structural member. The wide-flange steel beam economically responds to this basic principle. The member is designed in such a way as to put the most steel where it's needed, at the top and the bottom, to counter the bending stresses. The same principles apply to wood, but it's wasteful and expensive to cut wood in the wide-flange shape. Instead, wood is cut into rectangular sections of narrow width and large depth. Were timber to become as precious and costly a resource as steel, we might see lumber mills cutting more economically designed members. (Interestingly, some manufacturers are producing composite joists made out of plywood and lumber that closely resemble I-beams in shape.) Cross-Sectional Dimensions The cross section of a structural member refers to the shape and area of a cut perpendicular to its length. When we speak of cross-sectional area, we refer to the amount of material in the cross section. In a rectangular section this would be the width times the depth of the cross section. If we speak of the shape of the section, we are talking about its proportions. It is a fact that a rectangular member when placed on its edge (long side vertical) will be able to span a longer distance and support a larger load than if it was placed on its flat side. Actually, the strength of a structural member isn't so much dependent on the amount of material in the cross section as it's on the depth of the section. The most efficient section has most of its material (cross-sectional area) furthest away from its neutral axis. In doing this it's putting the most resistance where the stresses (compression and tension) are the greatest, at the top and the bottom of the section. Where the stresses are at their smallest (near the neutral axis), there is need for little material. A beam carrying a load across a span is subjected to a combination of stresses (or intensities of strain). The stress at the ends of the beam, where it's attached to the columns, is called shear. Shear stresses can be described as ripping in opposite directions. Bending stresses (, the combination of tension and compression, are greatest at the center of the beam span when the beam is uniformly loaded throughout.* It is important to know where these major stresses occur in a beam and how they affect its strength. As a general rule, you don't want to disturb the beam at a location where the stresses are greatest. For a uniformly loaded simple span beam, the shear stresses are greatest at the ends of the beam and zero at the center of the span, and the bending stresses are greatest at the center of the span and zero at the ends. For this reason, if you must drill a small hole through a beam, you should drill the opening along the neutral axis, toward the center of the span. In the design of joists, beams, and girders, the interrelationship among span, load, and the structural member’s cross-sectional dimension need to be evaluated to produce a satisfactory solution. *Bending stresses in continuous beams and in rigid frames are more complex. ROOF STRUCTURE A flat roof acts exactly like a floor and is structurally designed as a horizontal component system. Sloped roofs, on the other hand, present their own problems. If the roof is sloped (over 20 degrees), it begins to differ from a strictly load- transferring horizontal component. The sloped roof transfers the load diagonally. This diagonal thrust must be broken down into a vertical and horizontal component in order to design for its proper support. The vertical component is sup ported by the walls (or beams to columns). The horizontal component, or lateral thrust, must be accounted for as well; otherwise, there is a tendency for the roof to push the walls outward. This outward thrust could be accommodated by placing a buttress against the outer walls. That is the way the lateral thrust of the nave in the Gothic cathedrals was countered. Buttressing, however, is clumsy and unwieldy. The thrust could be countered by tying the rafter ends together using a steel chain or a piece of timber. This section is in tension—that is, it's being pulled or stretched. The most common solution in houses is to use a wood section as a collar beam or as ceiling joists to tie the rafter ends to each other, thereby countering the thrust. HOW IS YOUR BUILDING STRUCTURED? The following is a synopsis of the various building types. Your house or apartment building should fall into one of the categories. Apartment Buildings Most urban prewar apartment high rises are structured in steel. If you walk through one of the apartments in such a building, you will see a thickening of the wall in the corners of rooms in a rhythm (or module) of 12’ X 20’ or 20’ x 20’. The floor system is generally composed of steel beams and purlins (lightweight beams) spaced every 6’ to 8’ apart. In most apartment buildings constructed before World War II, this feature can be noted by examining the ceilings, which have plastered “drops” around the purlins. In these venerable old buildings, the interior partitions were usually constructed of 3”-, 4”-, or 5”-thick gypsum block, which is a lightweight, fire-resistant material that tends to crumble when cut into. Three coats of plaster—a brown, scratch, and finish coat*_were applied to the block wall. The partitions and , to some extent, the ceilings are solid. This means that you will have to channel into and then re-plaster the partitions should you add electric outlets, switches, or pipes. Most postwar apartment buildings were constructed of poured-in-place concrete beams and columns. The floor structure consists of a poured- concrete slab 6” or more thick. The finished flooring was applied to the top of the slab, and plaster (or in some cases just paint) was applied directly to the bottom of the slab as the ceiling. The partitions in these buildings are usually metal stud and gypsum board. The oldest urban apartment buildings, generally no taller than six to eight stories, have solid- masonry load-bearing walls and wood-joist floor systems. It is often difficult to determine the structure of the building without probing the ceilings and partitions. Buildings of this genre usually have some very thick partitions and some thinner ones. The wood joists generally span between the thicker partitions, which are load- bearing. In a load-bearing building, it's critical that you not accidentally remove a load-bearing partition or cut into the joists of the floor system. Wood-Frame Houses *The brown coat is composed of cement plaster, lime, and water. The scratch coat is similar to the first but is heavily scored (scratched) to receive the last coat. The plaster that's applied as a finish is called the white coat. Both the earliest and most modern wood-frame houses are based on a system called post-and- beam construction. This system is somewhat similar to the steel-frame model, using relatively heavy lumber at wide intervals to transfer loads from roof to foundations. Posts may be 4” x 4” or 4” X 6” in cross section and spaced 6’ to 8’ apart. Floor beams and roof rafters are often 4” x 12” or more in cross section and are spaced 4’ to 8’ apart. The floor and roof are structured with planks that are about 2” thick. In the early days of heavy timber construction, care was taken to ensure the rigidity of the heavy frame by providing diagonal bracing. Originally the connections were made with wooden pegs. Today, these connections are made with steel timber connectors; nails are just not strong enough to hold the connections for post-and- beam construction. The advantage of the fragile for a building, utilized a number of light weight wood sections spaced inches apart instead of heavy timbers spaced many feet from one an other. The predominant characteristic of the balloon frame is the use of 2 X 4 mini-columns (studs) that run vertically from the foundation to the roof. The frame isn't rigid in itself (it will not be able to stand up at right angles in a windstorm) but requires the addition of diagonal bracing or plywood sheathing to keep it from swaying under pressure. This method of construction is still post-and-beam frame today is to open the exterior walls for very wide windows and to allow for a feeling of openness on the interior. Light-frame construction differs from the heavier-frame systems in that it uses much thinner and lighter structural members, usually not more than 2” thick. These studs (2 X 4’s for the interior partitions and 2 X 6’s for the exterior walls), joists (2 X 10’s, 2 X 12’s), and rafters (about the same sizes as the joists) are spaced 12”, 16”, or 24” apart. The framing members are nailed together and are braced by applying ¾” plywood sheathing over the outside frame. This lightweight-framing method works almost like a load-bearing wall rather than a point support. Currently, two light-framing methods are used for residential construction: the balloon frame and the platform (or western) frame, the latter being the most popular of the two. The balloon frame was invented in the 1830’s and has been used for over 150 years. This system was designed in the Midwest just after the inventions of machine-made nails and the water- powered sawmill. The balloon frame revolutionized the house construction industry since it required only two or three workers to erect the frame and reduced its cost by almost 50 % . This system, disparagingly labeled “balloon” by its earliest critics, who thought the frame was too being used today for two-story houses faced with masonry or stucco, since there is less overall shrinkage in the frame that might cause cracks in the stucco. The easiest way to determine if your house is a balloon frame is to remove a small portion of the exterior siding in an area between the first and second floors. The studs are two stories long in a balloon frame, and the joists are nailed directly to them. If your house is two or more stories high and is finished in stucco, chances are the structure is balloon frame (Ills. 28 and 29). The simplest method of wood-frame construction is the platform frame. In this method, the builder constructs the entire subfloor before beginning the walls. In doing so, the worker has a flat and safe surface on which to build the wall sections. These sections are usually constructed horizontally on the floor with studs that are only one-story high and are then tilted up into place, temporarily braced, plumbed straight, and nailed to the subfloor. This system is the easiest to construct and will be the one discussed in the section on framing the extension. The interior partitions in all wood-frame construction consists of wood studs (although some builders are using metal studs now) covered with either plaster (in the old houses) or gypsum board (in houses built after World War II). In either case, you can pretty much depend on the partitions being hollow and easily modified. Most wood-frame houses, whether sided in wood or trimmed with a small amount of brick, are structured as platform frames. This method requires the least amount of cutting and is easiest to construct whether you are an experienced carpenter or a novice. Garden Apartments and Low-Rise Apartment Buildings Many apartment buildings located outside of the city and not more than three stories tall are built using platform-frame construction and sided with wood or with brick. The interior partitions are either wood or metal studs covered with either gypsum board or plaster. Some of the turn-of-, the-century walk-up urban buildings are structured with masonry load-bearing walls with wood-joist floor systems. Townhouses and Brownstones Most attached housing is based on the structural principle of the load-bearing party wall. Most older buildings have masonry party walls; newer buildings are framed in fine-rated metal stud assemblies. Walls between the units support the floors, consisting of wood beams (or joists) that span between the walls, supporting planks (or plywood) . The end walls of the house are generally non-load-bearing (except that they carry their own weight) and serve to separate the inside from the outside. If the townhouse is less than 16’ wide, it's likely that the beams that span between the party walls do so without any intermediary support.* The interior partitions may be composed of either studs, lath and plaster, or studs and gyp sum board, depending on when the building was built. *Some wide, old brownstones (say, over 19’ wide) have a “prop” wall, an intermediary supporting partition that runs parallel to the long side walls, even though the joists run continuously from party wall to party wall. If you are considering removing anything that may be a prop wall, be sure to consult an architect or structural engineer. Next: Structural Building Materials |
